Skip to content

High-performance browser automation bridge and multi-instance orchestrator with advanced stealth injection and real-time dashboard.

1.7 KB
Raw Blame

Onboarding persona × condition × goal matrix

Goal: pressure-test new-user onboarding. A blind agent simulates each persona reaching a first-step goal using only user-facing docs, in the persona's container.

ID Container Persona (skill) Starting condition First-step goal Primarily probes
S-A ptux-sa Impatient first-timer (low) chromium present open example.com, read title happy path + IDPI block for a non-technical user
S-B ptux-sb Backend dev (high) no browser scrape a page's heading no-browser cold start, scraping flow
S-C ptux-sc Privacy researcher (med) chromium present browse the public web IDPI allowlist / guards journey, security legibility
S-D ptux-sd DevOps / CI (high) no browser, minimal scriptable title (JSON/exit codes, restart) scriptability, restart cleanliness
S-E ptux-se Mis-set config (med) chromium + bogus browser.binary get nav working again diagnosability (doctor), recovery
S-F ptux-sf API integrator (high) chromium + curl open a page via the HTTP API API discoverability, auth/token UX

Per-scenario record (each agent returns)

Chronological friction log → for each step: command, key output (quoted), user interpretation, could-proceed-from-the-messaging-alone (yes/partial/no), one-line fix. Then: goal achieved? top friction points; onboarding verdict for that persona.

Synthesis (after a run)

  • Cross-persona friction table (recurring vs persona-specific).
  • Ranked onboarding improvements scoped to what's fixable in this repo.
  • Validation that prior fixes still hold under persona pressure.